Scheduling is not a workflow one normally associates with medical device connectivity. In some applications, scheduling is handled by software separate from the connectivity solution. Sometimes, scheduling is not done at all. In other applications, as we shall see, scheduling is so much a part of the broader workflow, that it's hard to recognize as a scheduling task. Two illustrative aspects of scheduling will be discussed, scheduling for diagnostic modalities and scheduling for routine patient care tasks. Because it's less understood (and frankly more interesting) we will look at scheduling for routine patient care tasks first.

Patient Care Task Scheduling

Patient care tasks encompass routine activities carried out by caregivers and/or aids. Examples of these routine tasks include vital signs collection, medication administration, bed turns (to avoid hospital acquired pressure ulcers, or HAPU), and respiratory circuit flushing (to avoid ventilator acquired pneumonia, VAP). These tasks must be completed at a predetermined frequency on a reliable basis or adverse events - including patient death - can result.

While the scheduling workflow diagnostic tests is very medical device centric (getting the patient to the device), patient care task scheduling is more patient centric (as in ensuring that certain patient care tasks are completed). The order for these routine tasks come from the ordering physician in numerous ways. Often the actual patient care task is implied by the physician order and must be interpreted by the caregiver. Some tasks are initiated based on operating policy that requires that patient's be screened for things like HAPU, VAP or fall risk. Patients that meet the at-risk criteria then receive the routine care prescribed by the policy.

Identifying all the routine tasks associated with a given patient is not the hard part. The challenge is ensuring that these routine tasks actually get done, and completed within the specified time frame. The reason this is a challenge is because of the interrupt driven environment at the point of care. This adverse work environment is at the root of many patient safety challenges found at the point of care: medication administration errors, hospital acquired pressure ulcers, ventilator acquired pneumonia, fall prevention, failure to rescue, and more. Besides failure to rescue, everything on the preceding list succeeds or fails based on completing routine tasks.

The need to balance nursing vigilance, medical device alarm response, patient and family member requests, and reliably completing routine tasks is a tall order. Past efforts to improve this situation have focused mainly on trying to develop and apply information technology to transform the point of care into a more manageable and predictable environment. Sadly, there is no information technology in existence that can direct when a patient needs to use the toilet, when their pain becomes intolerable, or when a patient's condition deteriorates generating a medical device alarm. Not surprisingly, attempts to reduce the interrupt driven nature of the point of care have failed.

While it is possible to somewhat improve improve workflow and nursing unit design to minimize interruptions, solutions that bring meaningful improvement to reliably completing routine tasks on schedule remain scarce.

Reliably completing routine tasks can be thought of as a connectivity solution. Let's consider bed turns as a means to prevent HAPU. First the patient must be screened and identified as at-risk, and an appropriate prevention regime selected - this portion of the workflow is well understood and widely adopted. What's missing for many point of care tasks is a solution that improves on the implementation of the previously selected patient care plan. In our HAPU example, the next requirement is a means to reliably know whether or not a patient has been turned within the prescribed time frame. Next, you need a means to prompt the caregiver to complete the required turn that doesn't itself become another nuisance interruption that detracts from patient care. Finally, the data from this process - when the the actual turns occurred compared to when they were scheduled - is recorded and available for retrospective analysis.

The first example of a solution supporting a specific type of routine care has recently come to market. A similar framework and resulting product could be used to address a variety of activities at the point of care. Some of these routine tasks are more challenging to support with automation than others. A perfect example of a challenging application is medication administration, where initial solutions were shown to be inadequate.

Diagnostic Scheduling

A classic example of scheduling is found in diagnostic imaging, where complex algorithms are needed to match requests for specific diagnostic procedures with available diagnostic equipment/rooms and human resources. This scheduling must be done in a way that evenly disburses workload across on-duty techs and radiologists and also maximizes the utilization of fixed assets such as x-ray rooms, CTs, MRs and interventional radiology suites. These kinds of complex scheduling tasks are often automated using software separate from the connectivity solution - in this example, scheduling is typically found in the Radiology Information System rather than the PACS.

These classic scheduling requirements are common to many diagnostic departments and associated with diagnostic connectivity workflows. Smaller, lower volume diagnostic modalities such as endoscopy and the cath lab may have scheduling included in the same solution as medical device connectivity - either from the medical device manufacturer or a third party.

If there was a connectivity solution for dialysis, this would be an example of a therapeutic modality where patient flow and resource utilization is as important as with many diagnostic modalities. An application like this is simplified in that the dialysis therapy does not vary like different diagnostic imaging procedures, and the dialysis machines on a unit tend to be identical or at least very similar. This greater degree of uniformity means the scheduling process is much less complex.

In certain situations scheduling can morph into more of a workload optimization and fulfillment exercise, somewhat different from the typical prospective scheduling scenario. For example, orders for clinical lab tests are generated by providers on nursing units or in their offices. These orders are received by the lab information system (LIS) which then dispatches phlebotomists to collect specimens which will then be tested.

A key part of the scheduling process is the initial capture of patient information. These patient demographics are captured along with the specific exam being ordered, the ordering physician and any special instructions regarding the patient or required time frame for the study. Some of this data, such as orders, may be available from other systems and can be pulled in without requiring the user to reenter that data. In some workflows, the order may follow the request to schedule the study. When this happens, there must be a validation step where the scheduled study and the ordered study are compared to ensure they are identical - and the ability to resolve any inconsistencies found.

Besides the obvious value of scheduling the study or some other patient encounter, scheduling data can indirectly support operations. For example, the schedule can drive when to push work lists or copies of orders to medical devices and/or techs to improve workflow. Scheduling data can also be used to determine optimal staffing levels for the scheduled workload. The scheduling of tasks can be part of a connectivity workflow, or a point of systems integration that feeds data into the connectivity workflow.

The foregoing scheduling workflows are mostly well understood. The software for supporting these kinds of scheduling tasks is mature and  most of these markets have reached penetration and become replacement markets.

What Have We Learned?

Routine tasks in patient care can be approached as scheduling challenges. Unlike with scheduling diagnostic tests, patient care task scheduling is more about the implementation of the tasks than the determination of when and where they should occur. There is a rich set of patient care tasks at the point of care that present persistent challenges to consistent implementation, resulting in adverse events and subsequent attention from the Joint Commission, AHRQ, CMS and others. It seems the industry - manufacturers and providers - are just now starting to come to grips with these routine clinical tasks.

Tim Gee is Principal of Medical Connectivity Consulting. He is a master connectologist, technologist and strategist working for medical device and IT companies and various provider organizations. You can learn more about Tim here.